Sensory

 

Introduction

Sensory applies the Truly HandsfreeTM technology to the ThinQ.AI to provide various functions for trigger word recognition. Sensory's HandsfreeTM technology delivers the performance greater than the industry standards by 40% or more. ThinQ.AI provides trigger word recognition and speaker identification functions among Truly HandsfreeTM functions.

  • FT (Fixed Trigger)
  • EFT (Enrolled Fixed Trigger)
  • UDT (User Defined Trigger)
  • Speaker Identification

 

Engine Structure

The engine provided by Sensory is also designed in the same structure as ThinQ.AI's existing engine.

The Sensory engine takes voice data as an input and provides an output in JSON format. The output includes the trigger word detection result and speaker information.

You can select properties for the functions you need among FT, EFT, and UDT.

 

Sensory architecture.png

 

Examples of Use

The trigger word detection function can be used for all devices equipped with the speech recognition function. In situations where you cannot touch a device directly, you can perform an action you want via the speech recognition function of the device.

 

  • Detect trigger word

You can operate the smart speakers using the trigger word detection function.

 

sensory usecase-1

  • Distinguish the users

If you register each user's voice on a common device used by multiple users, you can distinguish users through the speaker identification function to provide customized services.

sensory usecase-2

  • User-defined trigger word

You can assign a nickname to a personal device by registering a trigger word you want, instead of a fixed trigger word.

 

sensory usecase-3